Sparse preconditioning for model predictive control
Optimization and Control
2016-10-20 v2 Systems and Control
Abstract
We propose fast O(N) preconditioning, where N is the number of gridpoints on the prediction horizon, for iterative solution of (non)-linear systems appearing in model predictive control methods such as forward-difference Newton-Krylov methods. The Continuation/GMRES method for nonlinear model predictive control, suggested by T. Ohtsuka in 2004, is a specific application of the Newton-Krylov method, which uses the GMRES iterative algorithm to solve a forward difference approximation of the optimality equations on every time step.
Cite
@article{arxiv.1512.00375,
title = {Sparse preconditioning for model predictive control},
author = {Andrew Knyazev and Alexander Malyshev},
journal= {arXiv preprint arXiv:1512.00375},
year = {2016}
}
Comments
6 pages, 5 figures, to appear in proceedings of the American Control Conference 2016, July 6-8, Boston, MA, USA. arXiv admin note: text overlap with arXiv:1509.02861